Key points
Broad listening is a method that uses AI to collect, analyze, and visualize the opinions of many people.
Shibuya City used it for the free-text responses in its resident awareness survey, and the Tokyo Metropolitan Government used it for the call for opinions on the 2050 Tokyo Strategy.
It is used not as a statistical opinion poll but as a way to show the structure of the opinions that were collected.
Updated
Overview
The Tokyo Metropolitan Government describes broad listening as "a method that uses AI technology to collect, analyze, and visualize the opinions and voices of many people." Its distinguishing feature is that it organizes a large volume of written input, such as free-text survey responses and calls for opinions, by issue so that it can be viewed as a whole.
Local governments already have several mechanisms for hearing residents' voices. However, the more opinions are collected in writing, the harder it becomes to read them all, and reports sometimes present only some of the voices. By organizing and presenting those voices as a whole, broad listening makes it easier to grasp what issues exist and where opinions diverge.
Below, we explain in turn the public cases in local governments, how broad listening differs from opinion polls, how to connect it to discussion (deliberation), and Plural Reality's involvement.
Public cases in local governments
The following two are cases that each local government has made public. They were not carried out by Plural Reality.
Shibuya City: Analyzing the free-text responses of its resident awareness survey
On July 2, 2025, Shibuya City announced that it had analyzed 6,037 responses in the free-text fields of its fiscal 2024 resident awareness survey with "Kouchou AI," developed by Digital Democracy 2030 (DD2030). It carried out automatic summarization, classification and visualization by topic, analysis of trends by generation and by area, and extraction of issues.
Source: Shibuya City press release (PR TIMES) (opens in a new tab)
Tokyo Metropolitan Government: Analyzing the call for opinions on the 2050 Tokyo Strategy
The Tokyo Metropolitan Government invited submissions on "The Tokyo of the 2050s as You See It" from November to December 2024 and collected 27,915 opinions. It analyzed and visualized the collected opinions with AI and used them to review the policy framework of the new strategy.
Source: Tokyo Metropolitan Government, "2050 Tokyo Strategy" pocket edition (opens in a new tab), Tokyo Metropolitan Government, broad listening analysis results site (opens in a new tab)
How it differs from opinion polls
The results of broad listening show the structure within the opinions that were collected. They are not a statistical opinion poll. The people who respond to a call for opinions tend to be those with strong interest or those who find it easy to give their opinions, and the results can be skewed toward their voices.
| Aspect | Broad listening | Opinion poll |
|---|---|---|
| Main purpose | Survey what issues and differences of opinion exist | Estimate proportions, such as for and against, across the whole |
| Voices collected | Free-text responses and opinions that are submitted | Responses from respondents selected by a designed method |
| How to read the results | As the structure of opinions | As proportions |
| Points to watch | Can be skewed toward people who find it easy to give their opinions | With multiple choice, reasons and circumstances are hard to see |
The two do not replace each other; you choose between them or combine them depending on what you want to know.
How to connect it to deliberation
What broad listening reveals goes only as far as "what opinions exist." The next step, "what to do," is decided by people through discussion. One example of how to proceed is as follows.
- Listen broadly. Collect many voices, including free-text responses, through surveys and calls for opinions.
- Organize the issues. Classify and summarize the opinions with AI and make visible the points where opinions agree and diverge. Go back to the original text to check what the summaries say.
- Narrow down the issues and discuss them. Focusing on the points of divergence, hold discussions in residents' assemblies, councils, and similar bodies. Participants use the organized issues as a guide and share their reasons.
- Report back the results. Report back to those who submitted opinions what was decided in the discussion and what was carried over.
There are also ways to use AI to organize issues during the discussion itself. In Plural Reality's Cartographer (Baisoku Kaigi), AI creates statements from a theme that ask for agreement or disagreement, and participants answer on their smartphones with a degree of confidence from "strongly agree" to "strongly disagree." The points of agreement, the points of conflict, and the points with many "don't know" answers become visible, and the next set of questions can be recreated based on a transcript of the discussion. In an online workshop of Digital Democracy 2030, deliberation was carried out using this two-round approach.
Plural Reality's involvement
Plural Reality's AI tools were used in "Yosano Future Forum" (subtitled "Yosano Future Forum x Broad Listening"), a forum for resident participation toward the formulation of the Third Yosano Town Comprehensive Plan in Yosano Town, Kyoto Prefecture. It is a case that tackled the difficulty of "translating" residents' diverse voices into policy.
In addition, Shunsuke Takagi and Shutaro Aoyama contributed to the book "Broad Listening: Across Politics, Local Government, and Business" (in Japanese; DD2030 Community, Impress). Chapter 8 covers the introduction of Cartographer into Ota City, Gunma Prefecture's "Residents Council" (residents' assemblies run by KOSO NIPPON).
Related case studies and articles
Using AI to Overcome the Challenge of Translating Residents’ Voices into Policy
A Yosano Town case that used Plural Reality's AI tools under the banner "Yosano Future Forum x Broad Listening."
Read articleWe Contributed to the Book "Broad Listening: Across Politics, Local Government, and Business"
An announcement that members of Plural Reality contributed to a book on broad listening.
Read articleOpening Up Participatory Policy Deliberation with Cartographer: A Field Report
A record of a practice in which two rounds of online deliberation were held based on the opinions collected.
Read articleFrequently asked questions
- How is broad listening different from an opinion poll?
- An opinion poll is a survey that designs how respondents are selected in order to estimate overall trends. Broad listening shows what issues and differences exist among the opinions that were collected. Because it can be skewed toward the voices of people who find it easy to give their opinions, it cannot replace a statistical opinion poll.
- Were the Shibuya City and Tokyo Metropolitan Government cases carried out by Plural Reality?
- No, they were not Plural Reality's work. Shibuya City used "Kouchou AI," developed by Digital Democracy 2030 (DD2030), and the Tokyo Metropolitan Government analyzed and visualized opinions with AI in formulating the 2050 Tokyo Strategy. This page introduces them as publicly available information.
- Do the results of broad listening become policy conclusions as they are?
- The results are material for a decision, not a conclusion. After looking at the structure of the opinions, people narrow down the issues, discuss them, and decide. For this reason, one approach is to combine it with deliberation such as residents' assemblies.
- How is Plural Reality involved in broad listening?
- Plural Reality's AI tools were used in Yosano Town, Kyoto Prefecture's "Yosano Future Forum." In addition, Shunsuke Takagi and Shutaro Aoyama contributed to the book "Broad Listening: Across Politics, Local Government, and Business" (in Japanese; DD2030 Community, Impress).
Talk to us about your project
From designing a meeting or survey to running it, we help you decide where to start. Please contact us.
Contact us